ML20082D039
| ML20082D039 | |
| Person / Time | |
|---|---|
| Site: | Shoreham File:Long Island Lighting Company icon.png |
| Issue date: | 11/18/1983 |
| From: | Herr P MASSACHUSETTS INSTITUTE OF TECHNOLOGY, CAMBRIDGE, SUFFOLK COUNTY, NY |
| To: | |
| Shared Package | |
| ML20082C880 | List: |
| References | |
| ISSUANCES-OL-3, NUDOCS 8311220299 | |
| Download: ML20082D039 (60) | |
Text
_ _ _ _ _ _ .
7___________
.: y a
UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION Before the Atomic Safety and Licensing Board
)
In the Matter of )
)
LONG ISLAND LIGHTING COMPANY ) Docket No. 50-322-OL-3
) (Emergency Planning)
(Shoreham Nuclear Power Plant, )
Unit 1) )
)
DIRECT TESTIMONY OF PHILIP B. HERR ON BEHALF OF SUFFOLK COUNTY REGARDING CONTENTIONS 65 AND 23.D
Background
Q. Please state your name.
A. Philip B. Herr.
Q. Please state your occupation and qualifications.
A. I am an Associate Professor of City Planning in the Department of Urban Studies and Planning, Massachusetts Institute of Technology. A resume describing my qualifications is attached (Attachment 1). In summary, I have more than two decades of practice and teaching in the field of planning, dealing with Luch issues as probabilistic and other approaches 8311220299 831118 PDR ADOCK 05000322 T PDR
.-_________________._______________J
- r to contingency analysis, demography, transportation, infrastructure, and the relationships between plans and actions and between government and the governed. Because much of my practice has been in storm-prone New England coastal areas, an-ticipating emergencies has been a recurrent topic of concern.
In recent years, I have had several occasions to apply my back-ground and experience to the specific subject of emergency planning for areas surrounding fixed nuclear facilities including Shoreham, Seabrook, Pilgrim, and Montague (a plant proposed in western Massachusetts, but not built).
As a planner I am familiar with several types of analyti-l cal techniques. For instance, especially in planning for radiological emergencies, it is necessary to understand the logic of traffic models, interpret their results and judge their appropriateness. I have evaluated traffic models often in my non-emergency planning work. I have also made such ef-forts in evaluating evacuation time estimates for the Pilgrim plant, the Seabrook plant, and now the Shoreham plant.
Q. What is the purpose of your testimony?
A. The purpose is to address Contentions 65 and 23.D which concern evacuation time estimates. Those contentions allege i
1
- v i
that for a number of reasons, which I will discuss below, LILCO's evacuation time estimates are inaccurate and unreliable because they underestimate evacuation times in the event of a radiological emergency at Shoreham. As a result they are not reliable tools for planners or decision-makers.
Q. Why in your opinion are reliable evacuation time estimates important?
A. Evacuation time estimates have three primary purposes.
First, evacuation time estimates are used in plan preparation i as input to design choices, such as evacuation routes, traffic control measures, and the configuration of planning zones.
Second, time estimates are used in plan execution as a key input into action choices, such as the choice between recom-I mending evacuation and sheltering as the appropriate protective action for the public. Finally, evacuation time estimates are important because they provide a means for reaching judgments regarding whether effective emergency preparedness, (i.e.,
preparedness which actually would protect most residents if there were a serious radiological emergency) is feasible and can be implemented for a particular nuclear plant site.
, _.,-,-y,, -
- r Reliance upon time estimates which prove to be inaccurate could lead to inappropriate design or operational choices being made, possibly resulting in greater population exposure to ra-diation than necessary -- conceivably greater exposure than had no planning been done at all. Over-optimistic estimates, such as LILCO's, could lead to a recommendation to evacuate under circumstances where, in fact, many persons in vehicles could be overtaken by the radioactive plume rather than escaping from it. The result could be greater exposure than if sheltering had been ordered. Furthermore, public compliance wi.th evacua-tion plans depends in part upon confidence in the competency of those plans. Time estimates which lack reliability and credi-bility can do much to undermine that confidence, potentially resulting in increased non-compliance and exposure.1/
Q. Have you reviewed LILCO's evacuation time estimates?-
A. Yes. I have reviewed the time estimates shown in Appendix A of the LILCO Plan2 / and in a supplemental study known as l
l l 1/ See, for example, Perry, Ronald W., "The Social Psychology ,
of Civil Defense," Lexington Books, Lexington, Mass (1982).
-2/ This review has focused upon Revision 1. Revision 2 was received too late to be reviewed fully prior to completion of this testimony. Nevertheless, Revision 2 is reflected herein where possible.
- r KLD-TM-77,3/ both of which were prepared by LILCO's consul-tant, KLD Associates. I have also examined the demand estimates, network characteristics, and evacuation manegement system on which the LILCO time estimates are predicated.
Q. What are your conclusions?
A. In brief, I conclude that the estimates are too unreliable to form a competent basis for evacuation planning or operations. Among other things, there are clear deficiencies in estimates of mobilization times, analysis of non-evacuating traffic, and analysis of accidents and breakdowns. Also, the likely traffic delays and confusion resulting from LILCO's pro-posed traffic control measures were wholly overlooked. Final-ly, LILCO has failed to include in its evacuation time estimates (as set forth in Appendix A) the effects of the evac-uation shados phenomenon which will cause substantially more demand to be made on evacuation routes, thus increasing evacua-tion times.
-3/ KLD Associates, Inc., " Estimated Evacuation Times for the Entire Population Within the Emergency Planning Zone for the Shoreham Nuclear Power Station, Considering the Effects of Uncontrolled Evacuation, Voluntary Evacuation, -
Inclement Weather and Accidents" [ hereinafter "KLD-TM-77"].
o i
Contentions 65.A and 65.B Q. What issues are raised in Contentions 65.A and 65.B?
A. Contention 65.A alleges that LILCO's estimates of mobili-zation times are far too low because they do not adequately account for some mobilization activities. Contention 65.B adds to that the consideration that traffic congestion was not con-sidered by LILCO in developing mobilization estimates. I agree with both contentions.
Q. What is meant by the term mobilization time and what activities does it include?
A. As used in this contention and in the testimony of the Suffolk County witnesses, mobilization time refers to the time period between tne first evacuation notice to the public and the point in time at which all members of the public leave their homes, places of work, or institutions to begin to evacu-ate the EPZ. It includes such activities as leaving work and returning home to pick up family members prior to beginning to evacuate. Similarly, it includes a parent traveling from home or from work to a school to pick up children prior to commenc-ing the evacuation trip.
I
- D l
The concept of mobilization time used in Contentions 65.A and 65.B should not be confused with KLD's more restrictive use of the term, which it defines as a 20-minute period between the recommendation to evacuate and the time the first evacuee commences the evacuation trip. (Appendix A, Rev. 2 at V-3).
After the 20-minute " mobilization" period assumed by KLD, it further assumes a 2-hour trip generation period which represents the points in time between the departure of the first automobile on its evacuation trip and the departure of the last automobile on its evacuation trip. (See, Appendix A at V-4). Thus, using the County's terms, KLD's assumed mobili-l
! zation time is 2-hours and 20 minutes.
Q. What are the components of mobilization?
A. Figure 1 hereto (modified from NUREG 0654) shows the sequence of events in an evacuation as depicted by a series of "S"-curves. The shaded portion represents the events I comprising mobilization. As Figure 1 demonstrates, mobiliza-tion can be divided into such functions as notification, prepa-ration to leave work, work-to-home travel, and at-home prepara-tion time.
, - g , - , , _ _ _ , , - - - - -- ,
O loo ,, ,,,,,,,,.. , ,__,, . ______ ,,,,, ,,_.,,-
~ *,,'
)
' , [ . ~ h [ ' 's ! , *d
- ,e***
. . . f h 05 . *. . . :/ :. . , , ' '
9o :: anova
- *".c'j'. . !^
- r, .)*'*
^u 0- **
- eca'usarcom'"8
- :/ ,* i ,/ ** f*** , i RR04oCA$f,. ;** tr ay(NG : : ::/ / *
, ,e
, eo N ""'$-- _; fi-. :_:-- :- t " ' ,;!j'- - _ /_f- - _ j// -j'-cv'j^g51"h, - - --
, . - y.
- i. ' .
t , , ,/ cuarto d i,I! ! .... / I
. :* / /I I i
- w __ _.a.:=: _ni_om:m '
- .oucas : m _:r/_ _/____//_i_/_
- 1 @i 't @ l 0 :!! /
- I
.!. i[J, 7.uro-o=% / ,f '/ sesoot cuite.c.
- .. al;g!
5 3a .
- /,0.utario / / cvacuarto i
- a. :,:
- y: - : / evacuarto f f f; 4o .-:_ :-
-: - l * - - - - -;/-
' j'- _ _ _, ' ,- ,
- 30
- !! :: . :icI :/ ,
e' ,*
- i i
- !; !!: ./ f*i f* e i I l20 _' /_',,**',-. _ _ _I_ _ _ _ _ _ _ -l-- - - -
-. :i[/ -f. .'j a: lh: f/ ,_7,#..,.7
,, i 2 . . ..
v . j. .-. . :, : :g:. /
- *< .f*
Io - . ,.-
- 1.-
- . . . h**
- ..*. *,*I i i
/* ~
- /: . ;,.*d,**..*s,e*. .,.** i .* i i x o
2-45 2-3o 2-45 3 3-IS 3-3o 3-45 4 4 'S 4-3o o o-IS o-30 o-45 1 5-IS l-3o 6-45 2 TIME FRCN INITIATION OF PUBLIC NOTIFICATION a Alerting 1 . - - - _ ; . ;; 2 PueUC INFoRMATioN ORoACCAST
- B Example of Additional Reporting Format for Time Estimates of Population Evacuation When Probability Distributions j, Are Used i
> l.'
Note: These curves are suggestive of a hypothetical 10-mile radius EPI.
Similar curves ca.n be developed for sub-areas of the. entire EPI.
The horizontal displacement of these curves along the time axis .
as well as the slope of the curves will vary depending upon the characteristics of the EPZ or sub-areas of the EPZ.
Modified by Herr (.s ha d ing) from NUREG 0654.
Figure 1 f
l o
i l l
l Notification time. This is the time required for members i
of the public to become aware of the emergency situation fol-
- lowing the initial broadcast of the emergency notice. It will vary depending on a number of circumstances. Those residents I of the EPZ who hear the sirens, who know what they mean, and j who immediately turn on their radios will quickly become aware of the emergency. Others may not hear the sirens for a variety of reasons - for instance, because they are sleeping, because q
they have a hearing impairment, or because they are residents but are temporarily outside of the EPZ. Still others may not be in a position to have ready access to a radio or television.
Loss of power to the sirens could also delay notification.
Preparation to leave work. While some office workers need only put down a pencil and leave, other workers will need to secure work in progress (e.g., construction workers), secure permission to leave (especially if working outside the EPZ),
return to the place of work in order to leave (e.g., bus driv-ers), and resolve role conflicts (e.g., emergency workers with families residing within the EPZ).1/ All these activities will take varying amounts of time.
4/ See, Suffolk County testimony on Contention 25.
s
. , - . - . . , - . , - , , - - - -- - - - ' - ~ - - - - ~ - - - - - r-'- ~-- - - - - - - - ' - -'-~,e - ~ - - + - - - - -
e i
Return from work. -Evidence indicates that most persons evacuate as family units. This means that persons at work will i
seek to return home in order to unite with other family members prior to actually beginning to evacuate. A survey of EPZ re-sidents conducted on behalf of LILCO S
/ indicates that a tnird of the members of households surveyed normally require more than 30 minutes to return home from work, and that nearly 10 percent commute by rail transit. To the extent that one's return home to. reunite with the family is delayed by long commuting time, or the need to catch a train, mobilization times would be significantly affected.
Return from school and other locations. Under the LILCO Plan all school pupils except those in the Shoreham-Hading River School District are expected to be returned to their homes by school bus rather than being evacuated directly. (Ap-pendix at IV-182). Even assuming such a proposal were feasible (see Suffolk County Testimony on Contentions 68-69; Testimony of Erikson, Johnson, and Cole and representatives of School Boards on Contention 25), that trip may constrain mobilization 5/ See, Memo from G.P. King to File 166 re EPZ 0-D survey, j August 16, 1983 (obtained from LILCO in discovery).
1 t
1
- e times for affected fa- ies who must wait for children to return nome. Even LILCO concedes (Appendix A at V-1) that an accident while school is in session would delay its assumed 2-hour " trip generation" time, although it does not quantify the length of the delays. Assembling other family members may in-volve travel from day care centers, recreation areas, shopping, and the like.
Other preparation. Once the family is united at home, certain tasks will be necessary to prepare for evacuation.
LILCO's public education brochure, " Emergency Procedures --
Shoreham Nuclear Power Station" (rev. 2), at page 8, provides a partial list of such tasks, which include packing sleeping bags and blankets, necessary medicines (which may require a trip to a drugstore), clothes "for several days," important papers, checkbooks and credit cards, food for infants or pets, ob-taining a portable radio, etc. In reality, this is only a partial list of the tasks necessary to prepare for leaving one's home. Performing preparatory tasks will take time and, for many, these tasks will involve a number of trips within the EPZ before the evacuation trip. Highest priority would logi-cally go to tasks essential to successful evacuation, such as servicing the automobile. Second priority would go to actions
j l
which cannot be performed once having left the home vicinity, such as assembling clothing and supplies within the house, going to a local bank, obtaining medicin as from druggists, and checking on the well-being of neighbors or relatives. A lower priority likely would attach to other supportive actions, such i
as purchase of food or other items for use following evacua-tion.
Q. hhy is it important to include accurate estimates of mobi-lization time in deriving overall time estimates?
A. Mobilization time is a substantial component of overall evacuation time. In a survey of estimates for 10-mile.evacua-tions at 52 plants (NUREG/CR-1856, Table 5, page 10), notifica-tion time alone was found to be one-third of the median general population evacuation time. Thus, underestimating moblization time results in underestimation of a significant portion of the overall evacuation time.
The importance of deriving accurate estimates of mobiliza-tion times is particularly evident when one examines KLD's time estimates. For evacuations of less than 10 miles, the KLD time estimates are dominated by the mobilization times. For exam-ple, the overwhelming majority of the time KLD predicts to l
l L
j l
evacuate a two-mile at a around the plant is spent mobilizing.
Let us look at a two-mile evacuation under case 12 (in which KLD's assumptions of a 2 hour2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> trip generation time was uti-lized) and under case 21 (a sensitivity analysis in which KLD extended the trip generation time to 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br />). The evacuation i
times for these cases are presented in Table 1. Note that the first two subcategories of this table (" Mobilization" and " Trip Generation") constitute the activities comprising the County's concept of mobilization (although, as noted, the County does not concur with LILCO that the maximum value of this period j
equals 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> and 20 minutes),
f J
l 4
I 1 - . _ . _ _ . _ -
Table 1 1
Evacuation of 2-Mile Area l l
Case 12 Case 21 2-hour loading 3-hour loading Basis Time % Time %
Mobilization KLD " Mobilization
- Time" 0-20 10 0-20 10 KLD Assumption Trip Generation ** 2-00 60 3-00 85 KLD Assumption Sub Total 2-20 70 3-20 95 Other Times 1-00 30 0-10 5 Analysis Total 3-20 100 3-30 100 Source: Appendix A, Table XIV at V-3.
- Defined as time between recommendation to evacuate and departure of first cars.
- Time over which all cars commence evacuation trip.
l l
As can be seen from Table 1 above, mobilization time (as
! defined by the County) comprises 70 percent of the evacuation time estimated by LILCO for case 12 and 96 percent of the evac-l uation time for case 21, and is composed entirely of ungrounded assumptions. Even in the case of evacuation of the 10-mile EPZ, LILCO's assumptions regarding pre-evacuation activity com-prise 47 percent of the overall estimated evacuation time.
Thus, one can see that to the extent that mobilization times are based on unwarranted assumptions, as are KLD's, they render inaccurate the evacuation time estimates as a whole. Mobiliza-tion times that are too short because of faulty assumptions or the failure to consider significant components will result in an underestimation of evacuation times.
Q. Has KLD accurately modeled mobilization time?
j A. No. KLD asserts that within 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> and 20 minutes of initiating notification, all residents of the EPZ will leave their homes and begin the evacuation trip. KLD bases this figure largely on assumptions that are without factual support.
Q. Please explain.
l 1
A. Let us look first at the notification component. Evacua-tion times, to be meaningful to decision-makers, must include every time-consuming factor from the time an evacuation warning is given to the population. One such factor is notification time; the amount of time it will take for the public to become aware of an evacuation warning once it is initially given an evacuation notice. LILCO apparently expects that this period of time occurs during (rather than being added to) the 20-minute " mobilization" time or other LILCO component estimates, judging from the totals in Appendix A, Table XV. and the language of Appendix A at V-7.5/ In Table XV, Appendix A 6/ The evacuation' time estimates in Table XIV and elsewhere are the elapsed times from the beginning of evacuation to the time when all vehicles have left their respective indicated areas. In all cases, it is assumed that there is a time lag (called " mobilization time") between the time that the advisory to evacuate is announced and the actual start of evacua-tion. This mobilization time is estimated at 20 minutes. Hence, the 20 minutes must be added to the tabulated figures to relate the evacuation time estimates to the time that the advisory to evacuate is announced.
Note that the estimated time for the trip-generation process, which is the time over which people leave their residences or other facility to begin the evacuation trip, extends over a two-hour period (see Table X). This process is also assumed to (Footnote cont'd next page)
I l
1 j does note that " notification" will take 15 minutes.
1
- Second, KLD assumes a " mobilization" period of 20 minutes, i which it defines as the amount of time between a recommendation to evacuate and the time that the first evacuee begins the evacuation trip. The discovery process revealed that this 20-minute period is based upon a telephone survey of EPZ residents conducted for KLD,7/ which found that the median commuting time
! from work to home for EPZ residents, under normal conditions was 20 minutes. There is no basis for using this data to define a mobilization period. Why median work-to-home travel 4
should be the basis for " mobilization" time has not been ade-quately explained by LILCO. Many other factors, such as those
! I described above, (i.e., workers needing to shut down work or return to central office to obtain cars) are essential to determining when people will begin to leave places of work.
Furthermore, there are other types of mobilization trips, such i
i (Footnote cont'd from previous page) cegin 20 minutes after the time that the
- advisory to evacuate is announced.
- (Emphasis added).
l Appendix A, Rev. 2 at V-7.
I 7/ See fn. 3.
s j .
__m -y-_ m. ,- . , - - --- .- . -.-
~.w.m, _
I 5
l l as school-to-home travel, shopping-to-home travel, etc., that LILCO has not considered in selecting a 20-minute "mobiliza-tion" time. All of these factors, however, have a bearing on l
l l when persons will leave their homes. Thus, there is no justi-1 l fication for KLD to base its " mobilization time" solely on me-dian work-to-home travel. (See also, Pigozzi Testimony).
j KLD next asserts that after the 20 minute so-called "mobi-
, lization," all persons in the EPZ will have begun to evacuate within an additional two hours. Yet, there is no justification i for such an assumption. As I mentioned earlier, there are a number of factors that will affect mobilization and consequent-ly trip generation frequencies. In my opinion, some of these will lead to a trip generation period that may extend beyond i
several hours rather than the two hour period arbitrarily as-i sumed by KLD. For instance, school-to-home travel, which i
LILCO's time estimates do not consider at all, is likely to 4
take well over two hours under existing early dismissal pro-grams. (See Appendix A, V-1; Suffolk County Testimony on Con-l tentions 68-69). Thus, many families waiting for children to i
come home will be unable to begin evacuation travel within the time assumed by LILCO.
l i
i i
Finally LILCO's mobilization times do not consider the effect of traffic congestion on pre-evacuation trips (work-to-home, school-to-home, etc.) made for the purpose of preparing for evacuation. There will be a large number of commuters who will seek to travel home, shoppers who will return home, trips generated to bring school children home, and other trips taken in preparation for evacuation. There thus could be as many, or perhaps even more, pre-evacuation trips as there will be evacu-ation trips. The resulting high level of congestion on the roads as workers, shoppers and others attempt to return home, collect necessary provisions, etc. could cause delays in individual mobilization times, thereby delaying overall mobili-zation times and evacuation times. Again, this factor is ig-nored.
In short, the LILCO/KLD estimates of the period of time within which most people will actually begin to evacuate are unreliable. They are based largely on erroneous judgments and assumptions. If a proper analysis were performed, the time period for evacuation to get underway (and hence the overall evacuacion time estimates) would increase substantially.
i Contention 65.C Q. What is the issue raised in Contention 65.C?
A. The LILCO Plan is prescriptive in nature. It not only prescribes which persons will evacuate and when, as all evacua-
! tion plans do, but it also prescribes exactly the direction they are to travel and how they are to get there. Streets are turned one-way and particular turn movements are prohibited at many intersections. The Plan seeks to enforce these prescribed movements through the use of LILCO traffic guides to direct and control traffic, aided by traffic control devices such as signs and traffic cones.
LILCO's prescribed routes are invariant: they remain constant over all scenarios, and under all conditions. For in-stance, the evacuation route for Zone B is the same for a 2-mile 90' evacuation as for a 10-mile 360' evacuation; the same whether one or more of the assigned evacuation routes is impaired by construction or a massive accident; the same summer or winter; and the same for all individualc from the same zone i
or subzone regardless of their personal insights, intuitions, or objectives.
_ _ - - - - - - $ iw*s i-r 9 7
Contention 65.C alleges that because LILCO chose this au-thoritarian and unrealistic approach to traffic management, the Plan itself will become a cause of additional congestion and delay. That congestion and delay is not accounted for in ,
LILCO's time estimates, adding to their inaccuracy and unreliability.
Contention 65.C.4 i
Q. What is the concern stated in Contention 65.C.4?
A. The concern is that evacuees will not obey LILCO's pre-scribed evacuation routes because they will not perceive them to be the safest way out of the EPZ. The KLD model has sup-posedly optimized evacuation times by prescribing routes to evacuees, and assumes 100 percent compliance with them. There-fore, deviation from those routes will result in much higher evacuation times than KLD has predicted.
Q. Do you agree with Contention 65.C.4?
l A. Yes. First, there are many circumstances under which pre-scribed routes will in fact not be the "best" way out for i
evacuees. To the extent that evacuees are aware of.that, the prescriptions will cause' conflict, confusion, and anxiety. The i
I e
e-. e -. w... - - . - - - , .,a -- . , - _ . . ,ea . , - - +,e--
reasons lie in the basic structure of the LILCO evacuation plan. That plan prescribes a matching of source nodes (where people live or work) and exit points (destinations) (see Table X and Exhibit 2 of Appendix A), and then prescribes routes linking those source nodes and destinations. Those matchings and routes are invariant for all evacuation scenarios, with no provisions for contingent circumstances.
The plan design of destinations and routes was based upon a scenario of full evacuation to 10 and only 10 miles, assuming full compliance with evacuation orders. Both destinations and route assignments were colected to " balance" loads on the net- ,
work under the conditions of that scenario. (See Appendix A at III-8). However, under many circumstances the assigned desti-nation and/or route will be demonstrably less expeditious for some or all evacuees than some alternative would be. For in-stance:
(1) Assigned routes and destinations may be suboptimal for any ordered evacuation for less than 360* or for more or less than 10 miles. KLD's time estimates are premised upon assigned destinations and routes which are designed to i
" optimize" evacuation times under computer simulation, but not 4
l r
I l
l
in reality (See Pigozzi Testimony). However, even assuming these destinations and routes are optimal, since the assigned routes and destinations are not varied under different evacua-i tion circumstances, they lose their collective "optimality" ,
4 under any scenario other than full 10-mile evacuation.
Consider, for example, a plume projected to travel southward, resulting in evacuation recommendations to the l
i central quadrant of the EPZ.(which is depicted in Figure 2 to this testimony). This is the evacuation modelled by KLD as j Scenario 9. (See Appendix A, Table II).
t Under Scenario 9, people in Zone B would be ordered to
! evacuate. Starting two miles from the. plant, residents of the northern part of-Zone B are ordered to travel first over a mile eastward toward the plant (toward the projected plume direc-tion), then to join traffic from Zones A, C,G, and H, all using the William Floyd Parkway southward (see Appendix A, Figures 9, 11, 15 and 16), paralleling the plume's direction.
'l They are then directed to complete the trip with evacuees from i
other evacuated zones (D,E,I,J,M,P) and the " shadow evacuees" from the North Fork by travelling west out of the EPZ on the Long Island Expressway. Evacuees from Zone B following the l
l i
, _ _ _ _ _m. - , - r , . _ , , , , , , _ . - _ _ , , , . . , _ _ _ _ _ , , . . , _ , _ _ ~ _ - - _ __ ._
O e
0 giWEB g g @
Alte m te I
- k;.kjf' c .se f a'. .
.'s'
- r .'~'* *N
., !;: ::: :ll '
55b *: 5., ,
' {5l_, ]$5 ::: 5:Yl55 o : :.-
- .: :!!)fl::':. :.
Prescribed / '
,5)
- 5 5' .
.! .f.
Route A ... ':!. , ::fi:i
- .iii 5- 55,b .. {,
- :!: i: ::
ll! :!.
Plume ' -
Tra.jectory I
Figure 2 .
mandated route would finally clear the area from which evacuation is ordered about 12 miles after beginning.
Many if not most of the residents of Zone B will perceive this route to be unsafe and unreasonable. As Exhibit 2 demon-strates, those residents could travel a half mile west and im-mediately be clear of the ordered evacuation area expected by LILCO to be safe from the projected plume. Clearly, the re-sidents of Zone B would, under these circumstances, see that their self-interest is not served by following orders. In fact, those orders place them in so much greatet jeopardy that they would be foolhardy to follow them.
This is not the only example of such misrouting. Analyses of routes for other zones produce similar results. It is the inevitable consequence of a plan " optimized" for one scenario but unchanged for others. 1 (2) Assigned routes and destinations may be suboptimal for those families whose ultimate destination is not well-served by their assigned destination. For example, a person living west of the plant in Zone B intending tc go to Port Jefferson would be unlikely to follow the LILCO prescribed route or the direction of the LILCO traffic guides, which I
i direct him south to the Long Island Expressway. (See Figure 2). Anather example is provided by those who, unlike the ma-
. jority of people, and despite the plan's designers, might intend to travel to a destination east of the plant. In some circumstances, especially if they live east of the plant, some evacuees may perceive that travelling east would in fact mini-mize time and plant proximity far better than any assigned southward or westward route. Thus, they will have strong reason to deviate from assigned routes taking them west.
l (3) Assigned routes and destinations may be suboptimal for those families whose balancing of the relative importance l of minimizing evacuation time and avoiding proximity to the plant during evacuation results in a choice of route different from that prescribed in the plan. The plan's assignment of destinations and routes is designed to serve twin objectives:
rapid removal from the EPZ, and avoidance of travelling close to the plant while evacuating. (Appendix A at IV-14). There is not a unique " optimal" salution to such a two-objective problem, but rather many solutions, depending upon the relative weight given each objective. For example, some of those living east of the plant in Zone E might believe that travelling west on Route 25A would be quicker and therefore safer than their
assigned eastward and then southward route, even though that due west route carries them closer to the plant. Others i residing east of the plant, who are more concerned about prox-imity to the plant than the plan's designers, might seek first to travel eastward on Sound Road before turning south in order to put distance quickly between themselves and the plant.
(4) Assigned routes and destinations may be suboptimal for families departing early in the evacuation. Different routes may be optimal at different times. Some households are assigned to routes or destinations which require relatively long travel but which are presumptively optimal because of con-gestion on shorter alternative routes. But, for those evacuees, travelling the shocter route will in fact be more expeditious until congestion on that route reduces speeds sub-stantially below those possible on the longer prescribed route.
(5) Assigned routes and destinations may be suboptimal for cases where the assigned route's capacity is diminished by accidents, breakdowns, construction, flooding, or other contin-gency. In such cases, the assigned route will not be the most expeditious for some or all of those assigned to it, but rath-er, just as in the case of accidents during commuting hours,
l 1
l I
some alternatives will prove faster. However, using an alternative will require departing from the LILCO-prescribed
- routes and disobeying the LILCO tr*.ffic guides and traffic I
control devices, creating conflict, uncertainty, confusion and stress.
In each of these five examples, the prescribed route and destination will not be the most expeditious one for some or all evacuees. It takes no complex modelling or traffic expertise to see that it is so. Many of the evacuees will per-ceive that to be true, thus creating conflict between the assigned evacuation route and destination and the evacuees' perceived self-interest. In other cases, the prescribed route .
! and destination may actually be the best, but won't be per-
) ceived as such because of evacuee misjudgment, fear, and stress, wnich could lead to preference for simpler though i longer routes, or preference for routes more familiar than these assigned (see generally Saegert Testimony; SCPD Testimo-ny). In these cases, too, there will be conflict between evac-uation c'rders and the evacuees' perceived self-interest. -
Q. Will LILCO's traffic guides be able to prevent deviation?
l l
t
A. No. Assuming for the moment that LILCO has authority to direct traffic, the traffic guides could not enforce their orders. If Zone B residents wish to travel west, not east, they will do so either by disregarding the traffic guides or by using back roads on which there are no traffic posts. That will slow travel, create confusion, and promote conflict. (See Pigozzi Testimony.) No conceivable number of traffic guides can enforce compliance. Indeed, in the event of confrontation, 4
even a police officer would be hardpressed to enforce compli-ance with orders perceived to be threatening to health, such as described earlier for Zone B. However, a LILCO employee, no matter how garbed, will find it vastly harder to enforce com-pliance. (See SCPD Testimony.)
Q. What will be the effect of deviation from prescribed routes?
A. In some cases, such as the Zone B/ central quadrant evacua-tion scenario discussed earlier, departing from prescribed routes may both shorten those evacuees' own evacuation times and lighten loads on overtaxed assigned evacuation routes, thus shortening evacuation time for others. In many of the other cases, however, where routes are optimal but are not perceived 1
l l
l e
to be (particularly in the case of evacuation of the entire 10-mile EPZ) deviation from assigned paths could lengthen evac-uation times substantially, overusing obvious routes and underutilizing less obvious ones. (See Pigozzi Testimony).
In all cases, there is a behavioral effect: the conflict, uncertainty, stress, hesitation, and even blockage which can result from confrontations between evacuees seeking alternative routes and LERO traffic guides seeking to enforce the Plan.
Certainly such conflict would add to evacuation times. (See Saegert Testimony; Pigozzi Testimony).
Contention 65.C.1 Q. What concern is expressed in Contention 65.C l?
A. Contention 65.C.1 concerns the issue of screening. Revi-sion 1 of Appendix A called for LILCO traffic guides directing traffic at posts within the EPZ to screen motorists seeking to deviate from prescribed paths. Those persons who, in the judg-ment of the LILCO traffic guides, had " good reason" to deviate would be permitted to do so. Revision 2 of LILCO's Plan, how-ever, appears to have dropped the screening function. (See, letter from Jessine Monaghan to John Birkenheier, November 4, l
l
\
l
1983). Rather, through the use of barriers, cones or signs, LILCO assumes that deviant traffic will be discouraged, but that it will nevertheless be permitted.
Q. Does Revision 2 eliminate the concern raised in Contention 65.C.l?
A. No. Indeed, Revision 2 raises further questions. First, LILCO has provided almost no information on the " discourage-ment" process. On the surface, however, it appears to be mere-ly a semantic change rather than a change having any substance.
Appendix A, Rev. 2 at IV-23 states that the term " movements to be discourged":
Represents movements which, if made, will move traffic toward Shoreham Nuclear Power Station and/or is counterflow to the major evacuation traffic flow.
The only other clue to how " discouragement" operates was offered during the deposition of Edward Lieberman of KLD, who described the procedure as follows:
At the periphery of the network, and at the other control points as well, there will be layouts of cones, in the street, I indicating those thoroughfares where traf-fic will be discouraged, by the pattern of the cones.
)
i l
l l
l l
As a matter of fact, such cone patterns are applied at almost all control l points, to provide some visual guidance to the traffic flow.
But all of the cone layouts are de-ployed in such a way as to provide enough room for the vehicle to make that movement, if that vehicle is motivated to do so.
i Lieberman Deposition at 253-254.
1 Consider for a moment the driver approaching an intersec-tion, intending to make a left turn, who is confronted by an array of traffic cones and a traffic guide directing him to go to the right. In some cases, the driver may make the left turn without hesitation. In other cases, however, the driver will pause at the intersection, either to ask permission to make a left turn or because he does not perceive or understand that according to LILCO's scheme, he may drive through the cones.
This hesitation will cause delay as traffic behind that motor-ist backs up. The delay could even require the intervention of LILCO's tr af fic guide, thus raising the possibility of conflict between driver and traffic guide.
2 Second, once one person begins to ignore the cones and drive through them, others will follow until the traffic
. control scheme breaks down. Thus, it is clear that by eliminating the formal screening of motorists and accommodating 1
a deviant paths, LILCO's Plan no longer can assume strict compliance with prescribed paths. Yet, as I have discussed above, and as Professor Pigozzi discusses in his testimony, KLD's time estimates do assume strict compliance. Therefore, the Plan and the time estimates are based on inconsistent as-sumptions, thus making the accuracy of those estimates even, i
more doubtful.
Contention 65.C.2 Q. What issue is raised in Contention 65.C.2?
A. Contention 65.C.2 concerns the fact that delays will occur due to conflicts, and possibly aggression, between LILCO traf-fic guides and motorists. LILCO's evacuation time estimates do not account for such delays.
t O. Do you agree with Contention 65.C.27 A. Yes. Professor Saegert and the Suffolk County Police Department address the likelihood of conflict and aggression between evacuees and LILCO personnel arising from LILCO's traf-fic control scheme. Conflicts may result when motorists are l
j directed to move in a manner contrary to the way they want to go. (See discussion infra, see also Saegert Testimony). To l
l
1 i
the extent that LILCO's attempts to direct traffic result in such conflicts, they will certainly lead to confusion and in-terruptions in flow, thus raising evacuation times. (See also Pigozzi Testimony). If a motorist stops his car to argue with a traffic guide, he may very well block or delay traffic. It will also result in diverting the traffic guide from his assigned function of directing traffic, thus leading to further delays.
1 Contention 65.C.3 l Q. What concern does Contention 65.C.3 address?
l A. As LILCO has acknowledged, its traffic guides do not have the authority to alter traffic signals. Therefore, while they are directing traffic at a particular traffic post, the traffic t
l signals at that post will also be operating -- possibly in
, conflict with the traffic guides' directions. Such conflicting l
l directions to motorists will cause confusion and cause a reduc-tion in traffic speed while the confusion is resolved. Indeed, such conflicting direction may lead to increased accidents at intersectians. LILCO has not accounted for the delays caused by their traffic guides directing traffic contrary to direc-tions givens by signal lights.
. o Q. Do you agree with Contention 65.C.3?
A. Yes. The DOT Manual On Uniform Traffic Control Devices 8/
cautions that improperly operated traffic signal installations will have the following adverse impacts:
"1. Excessive delay may be caused.
- 2. Disobedience of the signal indications is encouraged.
- 3. The use of less adequate routes may be induced in an attempt to avoid such signals.
- 4. Accident frequency (especially the rear-end type) can be significantly increased."
The operation of traffic signals while personnel in the street give contrary directions is a prime case of improper signal operation. (See SCPD testimony). The conflicting di-rections are bound to create ambiguity, hesitation, confusion, and hazard. It is a matter of common sense that one should not give motorists conflicting directions, especially in the con-fusing and stressful circumstances that will exist during an evacuation from a radiological emergency. With LILCO's traffic guides rather than policemen at control points, the case against creating' conflict is even stronger. The result of
-8/ U.S. Dept. of Transportation, Federal Highway Administra-tion, 1971, at 216.
a i !
{ directing traffic contrary to signals will be confusion among
] motorists who will be unsure of the proper movement. Traffic 1
l will slow down as motorists attempt to resolve the confusion,
( tnus causing delays which will translate into higher evacuation i
times than LILCO has predicted.
Contention 65.D 4
Q. What is the concern addressed by Contention 65.D?
1 j A. The road capacities employed in KLD's traffic simulations 1
presume full use of the road: no accidents, breakdowns, r e.n-ning out of gas, road construction, or other capacity-reducing i contingencies. (Appendix A at V-2). However, in a j large-scale movement of automobiles over an extensive network, i
such as would be involved in an evacuation of the Shoreham EPZ, l
1 j there will in fact be accidents, breakdowns, and other such l
j obstructions, each of which will reduce capacity and increase I evacuation times. I concur with the conclusion stated in Con-tention 65.D that there will be a considerable number of i
j incidents having an adverse impact on evacuation times. Omis-l l sion of those considerations renders the LILCO estimates seri-
! ously flawed.
i i
i J
f 1
i
Q. Is it possible to reflect such obstructions in evacuation ,
time estimates?
A. Yes, by a number of methods including the following:
- 1. Approximation analysis by route. The number of vehi-cle miles on a given route is determined and multiplied by an incident rate to produce an estimate of incidents for that route. This type of analysis was conducted recently by PRC Voorhees and discussed below. (See Polk Testimony).
- 2. Arbitrary assignment in space and time. Capacity re-ductions could be built into the modelling system to reflect these incidents, with the locations and timing selected by the operator on some basis or bases. For example, a " worst case" analysis would take the largest credible number of accidents and locate them in the most damaging spatial and temporal pattern which is credible, and a "best case" analysis could do the opposite. This would cover the range of accident situa-tions and thus provide an illustration of the sensitivity of evacuation times to variation in the timing and location of accident conditions. KLD, in a supplemental accident analysis 9/ discussed below, allocated incident times and 9/ KLD-TM-77 at 24-26.
l i l l
l l
l l
I location arbitrarily, but without documenting or justifying ei-ther the basis or intent of such allocations.
- 3. Randomized assignment, using a Monte Carlo or other randomizing technique to assign times, locations, and I severities probabilistically. This could be done repetitively.
The evacuation time results from those repeated simulations would then describe a distribution of likely eva:uation times, providing stochastic description of the time consequences of incidents.
Q. What is the basis for your conclusion that there will be substantial numbers of incidents resulting in reduced capacity during an evacuation?
i A. My conclusion is based on an analysis conducted by PRC Voorhees which found that during an evacuation of the 10-mile EPZ, there could be as many as 140 accidents on the evacuation network and 277 incidents of automobiles running out of gas (See Polk Testimony and Attachment 4 thereto.)
Q. What is your opinion of PRC Voorhees' study?
A. PRC Voorhees' findings are credible, but, I believe, con-servative. That is, they somewhat underestimate the number of l
1 I
- __ h
l l
accidents and other incidents likely to occur. That conservatism is due to the following factors:
- l. Incidents on streets other than the seven major out-bound routes have not been considered. It is likely that there will be accidents on routes other than those seven, resulting in further delays of evacuating traffic.
- 2. Vehicle-miles experienced for work-to-home or other pre-evacuation trips have not been considered. However, it is clear that accidents and breakdowns will occur on these trips, also, having severe impacts: key evacuation roads will be im-peded; and persons mobilizing to evacuate will be further delayed.
- 3. Vehicle-miles and incidents outside the EPZ but im-pacting it through " spill-back" have not been considered. If, for example, a serious accident occurs on the Long Island Ex-pressway just outside the EPZ to the west of the plant, traffic will be backed into the EPZ. This will adversely affect evacu-ation times within the EPZ.
- 4. Travel by voluntary evacuees from outside the EPZ but travelling within it has not been considered. These additional
-. - . _. = - . .
evacuees may have accidents or breakdowns within the EPZ, which will further impede evacuation traffic.
- 5. All sources of vehicle disability other than accidents and running out of fuel have not been considered. Mechanical difficulty, driver incapacity, disability caused by weather or road conditions, and other causes could add significantly to those totals.
- 6. With as many as 140 accidents (or more, given the analytic omissions listed above), at least some will occur at 4 intersections. At those locations, they may have double im-pact, delaying travel on two routes.
- 7. The delay caused by incidents may last far longer than the time necessary to remove the obstructing vehicles, due to the " ripple effect" of an accident. Given heavy traffic, an accident involving a delay of any significance will result in the formation of a queue. Once the accident is cleared, the queue does not immediately disappear, rather it begins to dis-charge cars at a lower rate than the normal road capacity. Any obstruction which creates a queue has the effect not only of blocking traffic before the accident is cleared, but also, while the queue persists, of reducing road capacity from its l -
l l
I 4
uninterrupted capacity (up to 2,000 vehicles per hour per lane on an expressway) to its departing rate capacity (1,500 to 1,700 vehicles per hour per lane) as the queue discharges.10/
The effect of the obstruction ends only when the queue it cre-ated has been dissipated, at which point full capacity is re-stored. Evidence from heavily-travelled expressways indicates that this effect typically extends for up to 3 1/2 hours. A vehicle physically blocking a 2,000 vehicle / hour lane for 20 minutes introduces a blockage delay of 667 vehicle-hours, but empirical analysis shows an actual delay of 3,000 vehicle-hours.11/ LILCO's estimates ignore these real-world facts.
- 8. I believe that the Voorhees report somewhat underestimates the impact of accidents on two-lane roads.
Voorhees predicted no substantial effect where one lane of a two lane road was blocked by an obstruction. This was based on the assumption that traffic flow will be moving largely in one 10/ Transportation Research Circular, No. 212, Transportation Research Board, NAS, Washington, D.C., January 1980 at 256.
11/ Goolsby, Merrell E., " Influence of Incidents on Freeway Quality of Service," Highway Research Record No. 349, National Academy of Sciences, c. 1972.
direction, leaving the oncoming lane free for vehicles to use to avoid the obstruction. However, moving out of the normal
- travel path into the oncoming lane to go around disabled vehi-1
{ cles will certainly cause vehicles to slow down, thus reducing road throughput. The magnitude of the effect depends upon lane i
width, visibility, traffic control, and volume of oncoming traffic, so is difficult to quantify. My judgment is that the effect will be substantial and result in increased evacuation times. Moreover, I disagree that traffic will be substantially unidirectional during an ecergency. As mentioned earlier, peo-ple will attempt to take a variety of paths out of the EPZ, in many cases resulting in two-way flow. Furthermore, during the mobilization period, where cars are both leaving the EPZ and attem'pting to return home, there will be significant two-way flow. Therefore, an accident at this phase of an evacuation could interfere with one or both flows.
Q. Has LILCO attempted to address the issue of the impact of of vehicle disabilities on evacuation times?
A. Yes, although not in the Plan, or in the time estimates contained in Appendix A. Rather, a supplemental study conducted by KLD (KLD-TM-77) purports to examine the issue.
That study, to my knowledge, is not reflected in Revision 2 of the Plan and, thus, it is not clear whether LILCO relies on it or not.
Q. What is your opinion of that study?
A. Its conclusions are not reliable. KLD-TM-77 concludes that there will be only four accidents during an evacuation of the 10-mile EPZ. That figure is too low to be credible. For example, on the 1983 Fourth of July weekend, the Suffolk County police tabulated 329 reportedl2/ accidents on the roads of the five western towns in Suffolk County,13/ or about four per hour, including hours when traffic and incidents were very few.
That experience directly conflicts with an estimate of four accidents for a full evacuation.
KLD appears to have obtained its low accident figures be-cause, although it relied on the same data source as Voorhees (The Transportation and Traffic Engineering Handbook, ITE (2nd ed.)), it ignored the velocity-specific accident information 12/ See SCPD Testimony, note 17.
13/ The five eastern towns of Suffolk County are not within the SCPD's jurisdiction and thus are not included in this tabulation.
i l
which Voorhees utilized. That is, Voorhees took into account that the rate of accidents is higher for low-speed automobile travel, while KLD did not. (See Polk Testimony). The result is that KLD grossly underestimates the number of accidents likely to appear on the evacuation network during a radiological emergency.
Q. What will be the consequence of road construction on evac-uation times?
A. Again, the result will be varying degrees of delay depend-ing on the location and nature of the work. Such work will re-duce road capacity unless it can be entirely cleared from the roadway. Reduced road capacity will prolong evacuation time.
Neither the KLD nor the Voorhees analysis took account of that effect, and to that extent, they both understate the time nec-essary for evacuation.
Construction will in some cases oblige evacuees to depart from their assigned routes onto detours (such as when a road is completely blocked), and in other cases, will make it apparent to evacuees that their assigned evacuation route, although open, will be subject to delays because of construction, whereas an alternative would not. In both cases, there will be 9
.-. ,y, _ -
r
. c anxiety, stress, indecision, and delay caused by the inflexibility of the LILCO evacuation approach of invariant route assignments.
Q. The Suffolk County Police Department has testified that several portions of the evacuation network do not have adequate shoulders. What is the effect of this fact on evacuation time estimates?
A. Shoulders are important as a " safety valve," either to store a disabled vehicle or to serve as an alternative travel lane when the normal travel lane is blocked by a disabled car, construction or other cause. They also provide access for emergency vehicles. As the Suffolk County Police Department has testified, however, there are many sections of roadway in the EPZ where adequate or usable shoulders do not exist or where shoulders are interrupted by bridge abutments and light posts. In addition, during an evacuation, shoulders may be interrupted by the presence of disabled vehicles where it is likely disabled vehicles will be " stored." That will exacer-bate the effect of accidents, breakdowns, and construction since the absence of shoulder space will make it more difficult i
to clear obstructions or route traffic around obstructions.
l
1 1
- Furthermore, it can be expected that some drivers will ,
- I seek to use the shoulders of a road as an additional lane.
But, at those points where travel on the shoulder will no I longer be possible because of bridge abutments, light posts, or disabled vehicles etc., such vehicles will be required to merge with vehicles in the right-hand travel lane. That sort of
! merging process in heavy congestion will require drivers to ac-commodate one another through velocity adjustments, resulting l in a reduction in capacity in that merging section below that which it would be if there were no merging vehicles.14/ The result, again, will be reduced flow and increased evacuation times.
Contention 65.E Q. What is the concern expressed in Contention 65.E?
A. Under the LILCO Plan, a large number of special vehicles such as buses and ambulances will be on the evacuation routes during an evacuation. LILCO calculates that about 580 bus trips will be necessary to bring those without cars to transfer poir':s (Appendix A at IV-86e), plus further trips from the 14/ See reference at n. 8.
l
transfer points to evacuation centers. About an equal number of bus trips will be required to transport school'childen ei-ther home or to an evacuation center.ll/ Each of those
" loaded" trips will require an empty trip to initiate the ef-fort. In addition, there will be about 1000 ambulance trips generated by the evacuation of special facilities. (See Appen-dix A at IV-191). Thus, about 3,000 special vehicle trips are apparently anticipated by LILCO. There is no evidence, howev-er, that these additional trips are accounted for in LILCO's time estimates.
Q. How does that omission have an impact on evacuation time estimates?
A. The first consequence is that approximately 3,000 trips are not included in calculating the number of trips generated during an evacuation. Since buses are larger and have poorer acceleration than passenger cars, they occupy more road capaci-ty per vehicle (anywhere from 1.6x to 12x).15/ In addition,
--15/ Based on the following calculation: (140,000 pop. x 15%
school pop.) t 40 pupils / bus trip.
16/ Transportation Research Circular, No. 212, Transportation Research Board, NAS, Washington, D.C., January 1980, pg.
175, Tables 2.4 and 2.5.
(
1
.)
the buses will be making frequent stops, in some cases on des-ignated evacuatio.: routes (see, e.'., Zone C), blocking the :
lane for at least some time. Thus, these 3,000 trips have the consequence of roughly 5,000 passenger car trips (1.6 equiva-lent passenger cars / bus x 3,000 bus and ambulance trips),
adding about 10 percent to the total volume of vehicles modeled. Therefore, trip generation is understated by 10 percent. (See Professor Pigozzi's Testimony for a discussion of the need for accurate estimates of trip generation).
Second, bus trips in many cases are quite different in di-rection from the bulk of evacuating routes, since they are on return trips, creating counterflow, or are headed for transfer points. Some may be crossing evacuation routes. Choosing at random, Figures 13 and 13.1 of Appendix A indicate that bus route E-1 goes east on North Wading River Road while residents are instructed to go west. One consequence of this routing is the necessity of keeping reverse direction travel lanes open.
A second is that bus turning movements are in conflict with through movements. A third consequence is to further diminish resident confidence in the assigned evacuation route imperative, with drivers saying, "if he can go that way, why can't I?"
l l
= _ _ _ ,_
4 Contention 65.F ,
l Q. What is the concern behind Contention 65.F?
A. The concern is that stress and anxiety induced by fear of being " trapped" in a radiological emergency will diminish driv-ing skills. This could lead to an increased risk of accidents, an increased likelihood of route deviation, and slower travel speeds. None of these factors is reflected in the KLD estimates.
Q. Explain how impeded information processing will increase evacuation times.
As Professor Saegert has testified, driver performance will suffer under stress. Gerlough and Huber, Traffic Flow Theory, Transportation Research Board Special Report 165, 1975,ll/ clearly describes driver behavior in relation to infor-mation processing capacity (Chapter 5). They present empirical findings that " subjects with high accident rates had a very low information processing capacity." If, as Dr. Saegert has 17/ Gerlough, Daniel L. and Matthew J. Huber, Traffic Flow Theory. Transportation Research Board Special Report 165, Washington, 1975 at 75.
i
i testified (see Saegert Testimony), the peculiar circumstances of a radiological emergency on Long Island will reduce informa-tion processing capacity, that in effect will make high accident rate drivers out of low accident rate ones, i
The rate of vehicles moving past a point on a road is a
! function of velocity and spacing. Spacing depends upon a driv-er's ability to judge distance and velocity differences, which l
both require processing capacity. Reduced capacity could be expected to cause drivers to protect their safety by reducing velocity (reducing the volume of information to be processed in a unit of time) and increasing buffer space between themselves and the car in front, thus increasing headways. Most drivers behave exactly this way under conditions such as adverse visi-bility. Reduced information processing has an analogous effect.18/ The result of an impeded ability to process informa-tion will thus be increased evacuation times. This phenomenon is commonly encountered by police themselves, in what they refer to as " rookie accidents" which occur when responding under stress. (See SCPD Testimony).
18/ See fn. 13.
l l
l l
7 Contention 65.H Q. What is the concern stated in Contention 65.H?
A. Under the LILCO Plan, LILCO intends to assigr personnel known as "Eva'cuation Route Spotters" to patrol the evacuation routes. Apparently, they are supposed to observe road conditions and report them back to the EOC. Six persons are assigned to this task. The concern in Contention 65.H is that the congested conditions existing during a radiological emer-gency will make this task impossible.
Q. Please explain.
A. Route spotters will certainly be able to spot congestion, since it will be endemic; however, they will be unable to do anything about it. First, the route spotter's own mobility will be limited, at best. Analyses by PRC Voorhees show queues in excess of 1,000 vehicles or four lane-miles on a single leg on Route 347 (over 2,000 vehicles), the Long Island Expressway, and the Sunrise Highway.19/ Once caught in such congestion, the route spotters are unlikely to be able to escape it. Since 19/ See Polk Testimony, Attachment 3.
4 6 they will not have the special status or officially-marked automobiles of police cars or even ambulances, they will have little success in breaking through the congestion.
Second, once congestion is observed, the LILCO Plan leaves little room for relief, other than calling for tow trucks (if they are available). Remember that the LILCO Plan does not call for deviation from routes to accommodate existing conditions. If the plan were more flexible, an advisory could be broadcast suggesting that drivers avoid, say, Route 347, be- i 4
cause it is congested. However, the Plan is keyed to a deter-i minate response: once Route 347 is assigned, that is the route to be taken by the evacuees. That inflexibility renders the l number of spotters and their mobility nearly moot since routes i
will not change according to traffic conditions.
i l In short, the route spotters will be largely ineffective l in expediting traffic flow.
i 4
Contention 23.D 4
Q. What is the concern expressed in Contention 23.D?
i
- A. As Drs. Johnson and Zeigler explain in their testimony, i
l the TMI experience and studies on Long Island reveal that in i
i, l
l I ._ . - - - . -
I the event of an order to evacuate some or all of the EPZ, a substantial number of people not ordered to evacuate will do so. Accommodating their movement will require highway capacity which otherwise could exclusively serve those who are told to evacuate. This lengthens the time required for evacuating J
those told to leave, increasing their risk. It also means that in some cases, voluntary evacuees will be moving from locations
- of relatively low risk to locations of much higher risk. This propensity for persons not ordered to evacuate to do so -- the
" evacuation shadow phenomenon" -- has not been considered by KLD in deriving the time estimates in Appendix A. The result i
is distorted evacuation destination and route assignment, as well as understated evacuation time estimates.
Q. Is the shadow phenomenon of special concern on Long Is-land?
A. Yes. First, the topology of the Shoreham vicinity presents a special problem. Should an accident result in in-terdiction of an area just larger than 10 miles radius, the East End of Long Island would be isolated, with all east-west
! routes cut off. For many people, the spectre of that occur-rence looms large. As a result, there is special concern by i
i l
_ , _ _ . , _ . _ , , _ 9 _ _ - . - , .
and for the East Enders, who therefore, may be motivated to attempt evacuation without orders. The topology of Long Island means that for most, evacuation would require moving west, toward the plant, before getting away from it.
Second, population densities grow sharply just west of the EPZ, so even a relatively small fraction of the population -
evacuating means a huge number, in absolute terms, on the roads. Indeed, the number of voluntary evacuees will likely dwarf the numbers of " legitimate" evacuees on certain roads.
(See Johnson and Ziegler Testimony on Contention 23). Conges- t tion to the west could cause traffic to " spill back" into the EPZ, thus delaying the exit of those within the EPZ and thereby lengthening evacuation times.
Third, evidence indicates low public confidence in' - -
LILCO.20/ That low level of confidence could easily be trans- '
A '
lated into a high level of concern over official underestimation of consequences, leading to high voluntary ' '
[
evacuation on Long Island.
20/ Social Data Analysts, " Attitudes'Toward Evacuation: - *
,, g Reactions of Long Island Residents to a Possible Accident , _
at the Shoreham Nuclear Power Plant," (June, 1982).
A u *
~
~
. ~
~ l J M H*-
_W
Q. Has LILCO conducted any study of the shadow phenomenon and its effect on time estimates?
A. Yes. In KLD-TM-77, LILCO's consultant, KLD, examined the shadow phenomenon. That analysis is subject to all the methodological criticisms we have previously raised regarding the basic analysis. It also contains additional faults.
A flawed assumption in KLD-TM-77 is that shadow evacuees will take twice as long for mobilization as other evacuees, with the only stated basis being absence of sirens. The longer mobilization time led KLD to conclude that voluntary evacuees from the East End would trail evacuees from the East. Given
' the dominance of electronic media as a method of communication,
- J 'i that'is a clearly untenable assumption. Therefore, it is un-likely that East End evacuees will " trail" EPZ evacuees.
Furthermore, KLD has not included demand caused by East End evacuees traveling over the Sunrise Highway in its time estimates. As Mr. Polk explains in detail, there is no justi-
.j .fication for failing to' consider traffic on the EPZ network f
^
just because it originates from outside the EPZ. In doing so, however, KLD has underestimated the shadow effect.
~
, f
- c n r , _
F 4 e 5
- 53 ~
A A y v
t #
'4 e
f
1 Finally, KLD also looked at the effect of removal of ;
j special traffic controls from the network. This " uncontrolled" I
run yielded slightly higher times, but it is important to note that the " uncontrolled" evacuation retained the original ,
assigned set of destinations for each evacuation zone. The only control removed was the use of devices to manipulate in-tersection use, lane use, and route choice in proceeding to the given destination. For example, Zone B evacuees (see Figure 2) are still presumed to be heading for an exit node to the south-west, not due west, so their trip is probably little changed by these changed assumptions. As noted earlier, many people will deviate from their assigned destinations, thus rendering KLD's i time estimates inaccurate and unreliable.
On the other hand, PRC Voorhees has estimated much more realistic evacuation times for an evacuation of the 10-mile EPZ taking the shadow phenomenon into account. As Mr. Polk has I
testified, those estimates (under normal conditions and no I
accidents or breakdowns) are about 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> for the winter pop-l ulation and 18 hours2.083333e-4 days <br />0.005 hours <br />2.97619e-5 weeks <br />6.849e-6 months <br /> for the summer population.
Given the many deficiencies the KLD analysis, my judgment is that the Voorhees' estimates are more realistic and indicate i
i v v n , ~ , - -
.e - -,
7 l
- n both the degree to which the feasibility of timeful evacuation is misjudged in the LILCO Plan, and the degree to which plan i preparation has been flawed by resting on unsupported assump-i tions.
+ .
l Q. Please summarize your conclusions.
A. Evacuation time estimates, in order to be useful to plan-nets and decision-makers, must account for the realities of an evacuation. To the extent that such estimates reflect only a
ideal conditions, they do not reflect reality, because ideal J
conditions will not prevail in an evacuation.
In my opinion, the evacuation time estimates upon which the LILCO Plan is premised do not reflect reality. Contrary to the assumptions which underlie those estimates, during an evacution delays due to accidents and breakdowns will occur; people not advised to evacuate will evacuate; many people will i
not follow LILCO-prescribed evacuation routes; and there will be conflicts with emergency personnel. In addition, the need to prepare for an evacuation will slow the process considerably. Only time estimates reflecting these realities can be considered accurate and thus reliable tools for those who must make crucial decisions in the event of a radiological
f 1
- w l
l l
l t
emergency. The LILCO time estimates are, therefore, inaccurate l and unreliable.
i
'l l
I i
l l
t
( . .
I i
l I
ATTACHMENT 1 i
l i
i
, . - _ ._. _ . _ _ _ _ _ _ , . . _ _ _ _ _ .~. ..
- - - .- .. _ --= _ . .
- .~ ,.
A
~
ATTACHMENT 1 PHILIP B. HERR
.. EDUCATION .
Massachusetts Institute of Technology, Masters in City Planning, J.C. Nichols Fellowship. -
Rensselaer Polytechnic Institute, Bachelor of Architecture,
' Tau Beta Pi, Sigma Xi honoraries, Thesis Prize.
CURRENTLY Associate Professor of City Planning, M.I.T., Department of Urban Studies and Planning. Courses and research in growth and land use planning, participation, coastal zone management, -
design, impact analysis.
Principal, Philip B. Herr and Associates, consultants in land use -
planning, development regulation, impact analysis, partici-patory design.
Registered Architect, Commonwealth of ifassachusetts.
4 Member, American Planning Association, Urban Land Institute. -
RESEARCH PARTICIPATION
- Development Impact Assessment, funded by Massachusetts Department i of Community Aff airs, through Herr Associates, 1975-1976, and l Rockefeller Foundation, through M.I.T. Design of methods for l local analyses of development consequences.. . Publication
- . .. .
Evaluating Development Impact, M. I.T. Laboratory for Archi-tecture and Planning, August, 1978, i Environmental Impact Assessment, funded by Rockefeller Foundation and others through M.I.T. Laboratory for Architecture and
'- Planning, 1976-1978 (with Lawrence Susskind and ' others) .
Studies of institutional considerations in assessing compre-
' - hensive consequences of infrastructure -systems design, -case.
study of coastal zone management.
Maine Development Strategy, funded by Rockefeller Brothers Founda-tion and Maine Bureau of Public Lands, through M.I.T. Depart-l ment of Urban' Studies and Planning, 1974 (with Lloyd Rodwi'n l and others). ' Design of an approach to utilization of state-owned lands through new organizational approaches. Publica-tion: Economic Development and Resource Conservation: A' Strategy for Maine.
[ l
~ -
1 l l PHILIP B. HERR -- Page 2
.. RESEARCH PARTICIPATION (continued) %
Cambridgeport/Ecologue, funded by U.S. Office of Education, Office of Environmental Education, and others, through M.I.T. Depart-
' ment of Urban Studies and Planning, 1969-1972 (with Stephen Carr and others). Development of innovative methods for
- enabling community residents to develop neighborhood plans.
- Publication
- article in Progressive Architecture, December, ,
1976.
Mobility for the Poor, funded by U.S. Department of HUD, through the M.I.T.-Harvard Joint Center for Urban Studies, 1968-1970 q (with Aaron Fleisher) . Analysis of travel patterns and dis-
,! abilities of the poor, and of possible remedies, based on survey data from Boston, Memphis, St. Louis, Milwaukee and Baltimore.
CONSULTING l
Participatory planning and design. Program design and technical assistance for a variety of New England towns and regional planning agencies, including Bourne, Edgartown, Franklin, -
Gloucester, Oak Bluffs, Rowe, Sharon; Sherborn, Sunderland, and Tisbury, Massachusetts; Hanover, New Hampshire; Cape Cod Planning and Economic Development- Commission.
Innovative development control. Techniques designed have included growth timing (Bourne, Falmouth, Franklin, Greenfield, Sandwich);
performance zoning (Clinton, Franklin County, Gay Head, Sand-wich); transfer of development rights (TDR) (Sunderland) ;
' . critical resource zoning (Sherborn, Sunderland); regional land use control (Franklin County, Martha's Vineyard Commission).
Other development control. Over twenty zoning bylaws and ordinances have been rewritten and adopted, numerous other controls de-i signed and adopted in more incremental fashion.
- Impact analyses. Cape Cod National Seashore (for National Park Service), open space acquisition (for Association for Preser-vation of Cape Cod), dog track (for Blackstone), PUD (for i Natick), resort development (for Franklin County), nuclear
-power plant ~(for Franklin County).
l Central area studies. Amherst, Andover, Gloucester, Lexington, l Northampton, Salem, among others, in each case utilizing
! alternatives to conventional federal-aided urban renewal.
l Regional efforts have included "208" Water Quality Management plan-ning for Cape Cod, creation of a regional housing authority and !
regional building inspection system for Franklin County, model cluster zoning legislation for Cape Cod.
l
.. .pm-PHILIP B. HERR -- Page 3 JOURNAL PUBLICATIONS ,,
American Institute of Planners, Planners Notebook, October, 1973,
" Performance Zoning: The Small Town of Gay Head, Massachusetts, Tries It", with Kevin Lynch.
Eno Foundation, Traffic Quarterly, April, 1962, " Timing of Highway Impact".
Urban La'nd Institute, Urban Land, February, 1960, " Regional Impact of Highways".
Descriptions of Herr's community work have appedred in Progressive Architecture, November and December, 1976; Journal of the American Institute of Planners, January, 1975; Journal of Housing, May, 1980; Land Use Law & Zoning Digest, March, 1980; PAS Memo, March, 1980; The Land Use Controversy in Massachusetts (L. Susskind, Ed., 1975); Performance Stan-dards: A Technique for Controlling Land Use, Oregon State University Extension Service.
PR'EVIOUS EXPERIENCE Chairman, Planning Subcommittee, Massachusetts Governor's Task Force on Coastal Resources.
Member, Steering Committee, Massachusetts Coastal Zone Management Program.
Director of Planning (subsequently, President), Economic Development Associates, Inc., Boston 3 Mass.
Research Associate, Greater Boston Economic Study Committee.
Consulting Associate, Adams, Howard and Greeley, Cambridge, Mass.
Planner, City of Berkeley, California.
1
.. Instructor, Boston University, Wentworth Institute, Boston Archi-tectural Center.
Architectural draf tsman/ designer, George W.W. Brewster, Warren C.
Obes, Boston, Mass.
Member, Revere Beach Design Review Board (appointed by Secretary of Environmental Affairs).
.